| Biomedical Engineering |
|
|
|
|
| Evaluation of machine learning classifiers for diagnosing mediastinal lymph node metastasis of lung cancer from PET/CT images |
| WANG Hong-kai1, CHEN Zhong-hua1, ZHOU Zong-wei2, LI Ying-ci3, LU Pei-ou3, WANG Wen-zhi3, LIU Wan-yu4, YU Li-juan3 |
1. Department of Biomedical Engineering, Dalian University of Technology, Dalian 116024, China;
2. Department of Biomedical Informatics and the College of Health Solutions, Arizona State University, Scottsdale 85259, US;
3. Center of PET/CT, The Affiliated Tumor Hospital of Harbin Medical University, Harbin 150081, China;
4. HIT-INSA Sino French Research Centre for Biomedical Imaging, Harbin Institute of Technology, Harbin 150001, China |
|
|
|
Abstract The classification performance in diagnosing mediastinal lymph node metastasis of non-small cell lung cancer (NSCLC) was evaluated from (18F-FDG) PET/CT images with four mainstream classical machine-learning classifiers (random forest, support vector machines, adaptive boosting, and back-propagation artificial neural network). 1397 lymph nodes were segmented from 168 patients' PET/CT images, and 13 kinds of image features (Dshort, area, volume, HUmean(2D or 3D), HUcontrast(2D or 3D), SUVmean(2D or 3D), SUVmax(2D or 3D), SUVstd(2D or 3D)) were extracted from each lymph node. The extracted 13 kinds of image features were combined to get 4 kinds of combinatorial variables ("All features", "High AUC features", "Doctor's features", "3D features"). The diagnostic performance of random forest, support vector machines, adaptive boosting, and backpropagation artificial neural networks were quantitatively evaluated according to the four kinds of combinatorial variables in terms of sensitivity, specificity and area under the ROC curve (AUCROC). The evaluation results show that the four classifiers yielded sensitivity are between 77%-84%, specificity between 81%-84% and AUCROC between 0.86-0.90. Under the significant contrast conditions (p<0.001), although the specificity of machine learning methods is slightly lower than that of human experts, but the sensitivity is significantly better than that of human experts. Results showed that 3D features and PET-CT combined features resulted in significant improvement of AUCROC. Although the 4 kinds of machine learning methods demonstrate promising sensitivities for mediastinal lymph node metastasis of non-small cell lung cancer diagnosis from (18F-FDG) PET/CT images, their specificities still need to be improved. A variety of classification methods are needed to conduct joint experiments in the future, and more advanced machine learning methods such as deep learning will be used for the further study.
|
|
Received: 04 February 2017
|
|
|
机器学习算法诊断PET/CT纵膈淋巴结性能评估
评估4种主流典型的机器学习方法（随机森林、支持向量机、AdaBoost、反向传播人工神经网络）对（18F-FDG） PET/CT影像中非小细胞肺癌纵膈淋巴结良恶性进行诊断分类的性能.先从168例病人的PET/CT影像中分割出1 397个淋巴结，对每个淋巴结提取出13种图像特征（Dshort、area、volume、HUmean（2D or 3D）、HUcontrast（2D or 3D）、SUVmean（2D or 3D）、SUVmax（2D or 3D）、SUVstd（2D or 3D））；将提取出的13种图像特征进行组合，得到4种组合变量（“All features”、“High AUC features”、“Doctor's features”、“3D features”）；在4种组合变量下，分别从敏感性、特异性以及ROC曲线下的区域面积（AUCROC）3个方面对随机森林、支持向量机、AdaBoost、反向传播人工神经网络定量地进行诊断性能评估.评估结果显示，4种分类器分割结果的敏感性为77%~84%，特异性为81%~84%，AUCROC为0.86~0.90.在显著性（p<0.001）条件下对比发现，虽然机器学习方法的特异性略低于人类专家，但是敏感性显著优于人类专家.研究结果表明，三维图像特征及PET/CT影像组合特征可以显著提高AUCROC.基于上述研究结果可以得出结论，虽然4种机器学习方法在（18F-FDG） PET/CT影像的非小细胞肺癌纵膈淋巴结的良恶性诊断中展现了不错的敏感性，但它们的特异性有待进一步提高，在未来需要尝试多种分类方法进行联合实验，使用更高级的机器学习方法如深度学习进行进一步的研究.
|
|
| [1] 齐守良, 岳勇, 辛军, 等. 面向临床肿瘤诊疗决策的多模态医学影像融合[J]. 中国生物医学工程学报, 2013, 32(3):356-362. QI Shou-liang, YUE Yong, XIN Jun, et al. Fusion of multi-modality medical imaging information for clinical decision in tumor diagnosis and treatment[J]. Chinese Journal of Biomedical Engineering, 2013, 32(3):356-362.
[2] MCFIELD D, BAUER T. A review of noninvasive staging of the mediastinum for non-small cell lung carcinoma[J]. Surgical Oncology Clinics of North America, 2011, 20(4):681-690.
[3] SILVESTRI G A, GOULD M K, MARGOLIS M L, et al. Noninvasive staging of non-small cell lung cancer:ACCP evidenced-based clinical practice guidelines (2nd edition)[J]. Chest, 2007, 132(3):178S-201S.
[4] BRODERICK S R, MEYERS B F. PET staging of mediastinal lymph nodes in thoracic oncology[J]. Thoracic Surgery Clinics, 2012, 22(2):161-166.
[5] KIM S K, ALLEN-AUERBACH M, GOLDIN J, et al. Accuracy of PET/CT in characterization of solitary pulmonary lesions[J]. Journal of Nuclear Medicine, 2007, 48(2):214-220.
[6] LI Xiao-lin, ZHANG Hua-qi, XING Li-gang, et al. Mediastinal lymph nodes staging by 18F-FDG PET/CT for early stage non-small cell lung cancer:a multicenter study[J]. Radiotherapy and Oncology, 2012, 102(2):246-250.
[7] SCHMIDT-HANSEN M, BALDWIN D R, HASLER E, et al. PET-CT for assessing mediastinal lymph node involvement in patients with suspected resectable non-small cell lung cancer[J]. Cochrane Database of Systematic Reviews, 2014, 11(11):CD009519.
[8] DE-LEYN P, VANSTEENKISTE J, CUYPERS P, et al. Role of cervical mediastinoscopy in staging of non-small cell lung cancer without enlarged mediastinal lymph nodes on CT scan[J]. European Journal of Cardio-Thoracic Surgery, 1997, 12(5):706-712.
[9] ZHENG Yuan-da, SUN Xiao-jiang, WANG Jian, et al. FDG-PET/CT imaging for tumor staging and definition of tumor volumes in radiation treatment planning in non-small cell lung cancer[J]. Oncology Letters, 2014, 7(4):1015-1020.
[10] EDET-SANSON A, DUBRAY B, DOYEUX K, et al. Serial assessment of FDG-PET FDG uptake and functional volume during radiotherapy (RT) in patients with non-small cell lung cancer (NSCLC)[J]. Radiotherapy and Oncology, 2012, 102(2):251-257.
[11] WANG J, WELCH K, WANG L, et al. Negative predictive value of positron emission tomography and computed tomography for stage T1-2N0 non-small-cell lung cancer:a meta-analysis[J]. Clinical Lung Cancer, 2012, 13(2):81-89.
[12] LIAO Chi-ying, CHEN Jin-hua, LIANG Jian, et al. Meta-analysis study of lymph node staging by 18 F-FDG PET/CT scan in non-small cell lung cancer:comparison of TB and non-TB endemic regions[J]. European Journal of Radiology, 2012, 81(11):3518-3523.
[13] AMBROSINI V, FANTI S, CHENGAZI V U, et al. Diagnostic accuracy of FDG PET/CT in mediastinal lymph nodes from lung cancer[J]. European Journal of Radiology, 2014, 83(8):1301-1302.
[14] SILVESTRI G A, GONZALEZ A V, JANTZ M A, et al. Methods for staging non-small cell lung cancer:diagnosis and management of lung cancer, 3rd ed:American College of Chest Physicians evidence-based clinical practice guidelines[J]. Chest, 2013, 143(5):e211S-250S.
[15] 王伟胜, 骆嘉伟, 林红利. 医学图像计算机辅助诊断数据平台研究[J]. 中国生物医学工程学报, 2013, 32(1):105-108. WANG Wei-sheng, LUO Jia-wei, LIN Hong-li. Computer-aided diagnosis data platform by using medical imaging[J]. Chinese Journal of Biomedical Engineering, 2013, 32(1):105-108.
[16] WAUGH S A, PURDIE C A, JORDAN L B, et al. Magnetic resonance imaging texture analysis classification of primary breast cancer[J]. European Radiology, 2016, 26(2):322-330.
[17] JACOBS C, VAN RIKXOORT E M, MURPHY K, et al. Computer-aided detection of pulmonary nodules:a comparative study using the public LIDC/IDRI database[J]. European Radiology, 2016, 26(7):2139-2147.
[18] PARK H B, LEE B K, SHIN S, et al. Clinical feasibility of 3D automated coronary atherosclerotic plaque quantification algorithm on coronary computed tomography angiography:comparison with intravascular ultrasound[J]. European Radiology, 2015, 25(10):3073-3083.
[19] BENNDORF M, KOTTER E, LANGER M, et al. Development of an online, publicly accessible naive Bayesian decision support tool for mammographic mass lesions based on the American College of Radiology (ACR) BI-RADS lexicon[J]. European Radiology, 2015, 25(6):1768-1775.
[20] FEULNER J, ZHOU S K, HAMMON M, et al. Lymph node detection and segmentation in chest CT data using discriminative learning and a spatial prior[J]. Medical Image Analysis, 2013, 17(2):254-270.
[21] ROTH H, LU Le, LIU Jia-min, et al. Improving computer-aided detection using convolutional neural networks and random view aggregation[J]. IEEE Transactions on Medical Imaging, 2016, 35(5):1170-1181.
[22] KERHET A, SMALL C, QUON H, et al. Application of machine learning methodology for PET-based definition of lung cancer[J]. Radiotherapy and Oncology, 2009, 92(1):41-47.
[23] KERHET A, SMALL C, QUON H, et al. Segmentation of lung tumours in positron emission tomography scans:a machine learning approach[C]//Conference on Artificial Intelligence in Medicine:Artificial Intelligence in Medicine.[S.l.]:Springer, 2009:146-155.
[24] CHEEBSUMON P, BOELLAARD R, DE RUYSSCHER D, et al. Assessment of tumour size in PET/CT lung cancer studies:PET-and CT-based methods compared to pathology[J]. Ejnmmi Research, 2012, 2(1):1-9.
[25] LAMBIN P, DEHING-OBERIJE C, PERSOON L, et al. Machine learning based clinical research:the example of lung cancer[J]. Medical Physics, 2008, 35(6):2900.
[26] GAO Xuan, CHU Chun-yu, LI Ying-ci, et al. The method and efficacy of support vector machine classifiers based on texture features and multi-resolution histogram from (18)F-FDG PET-CT images for the evaluation of mediastinal lymph nodes in patients with lung cancer[J]. European Journal of Radiology, 2015, 84(2):312-317.
[27] ZHOU S, CHENG Y, TAMURA S. Automated lung segmentation and smoothing techniques for inclusion of juxtapleural nodules and pulmonary vessels on chest CT images[J].Biomedical Signal Processing and Control, 2014, 13(5):62-70.
[28] SHI C, CHENG Y, WANG J, et al. Low-rank and sparse decomposition based shape model and probabilistic atlas for automatic pathological organ segmentation[J].Medical Image Analysis, 2017, 5(38):30-49.
[29] LIU J, HOFFMAN J, ZHAO J, et al. Mediastinal lymph node detection and station mapping on chest CT using spatial priors and random forest[J].Medical Physics, 2016, 43(7):4362-4374.
[30] VALLIERES E, SHEPHERD F A, CROWLEY J, et al. The IASLC lung cancer staging project:proposals regarding the relevance of TNM in the pathologic staging of small cell lung cancer in the forthcoming (seventh) edition of the TNM classification for lung cancer[J]. Journal of Thoracic Oncology, 2009, 4(9):1049-1059.
[31] FRANCIS T. Machine learning:an algorithmic perspective, second edition (eBook)-Taylor&Francis[EB/OL]. https://www.crcpress.com/Machine-Learning-An-Algorithmic-Perspective-Second-Edition/Marsland/p/book/9781466583283. |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
